Colloquium: Rodrigo Ranero

Speaker: Rodrigo Ranero (UCLA)

Title: A new perspective on the syntax of silence: The view from Mayan

Abstract:

Ellipsis is structure and meaning without form. In the case of spoken languages, it is silence that requires a linguistic antecedent. An unresolved question concerns the precise nature of the relationship that must hold between the silence and its antecedent—the identity condition underpinning ellipsis. In a nutshell: Is syntactic, semantic, or some sort of hybrid identity required?

Data that serve to address this question involve cases where there is a mismatch between the silence and the antecedent, yet the result is nevertheless well-formed:

(1) Every night, Bianca lip syncs. Trixie taught her how <to lip sync>.

Of equal interest, there are other cases where there is a mismatch but the result is ill-formed. Voice mismatches in sluicing—clausal ellipsis with a wh-remnant—are a prominent example:

(2) *Bianca was crowned, but we can’t remember who <crowned her>.

Any explanatory approach to the identity condition must derive this asymmetry within individual languages like English, as well as any variation we might find cross-linguistically.

In this talk I will argue for a new approach to the identity condition. I propose that the condition incorporates a syntactic component, but strict identity is not required (contra a long tradition; see recently Merchant 2013, Rudin 2019). Rather, the identity condition requires that the ellipsis site and the antecedent be non-distinct:

(3) Proposal: syntactic identity in ellipsis
The antecedent and material properly contained in the ellipsis site must be featurally nondistinct.

In support of (3), we will discuss evidence from Kaqchikel, a Mayan language of Guatemala that possesses a rich voice system (García Matzar & Rodríguez Guaján 1997). In contrast to languages of the English kind, a subset of voice mismatches is well-formed in Kaqchikel sluicing. In particular, the Agent Focus voice—a voice that is specific to several Mayan languages— can mismatch with active and passive voices. To exemplify, a well-formed Agent-Focus-active mismatch is shown below:

(4) Xaxe ri ma Pedro x-Ø-loq’-o ri kotz’i’j.Aw-etam-an ankuchi
only DET CLF Pedro COM-ABS3S-buy-AGENT.F DET flowers ERG2S-know-PERF where

<x-Ø-u-löq’ wi>?
‘Only Pedro bought the flowers. Do you know where?’

I will argue that my novel identity condition derives the well-formed status of these hitherto undiscussed data and examples like (1), and it also derives the ill-formed status of examples like (2). The proposal thus brings us closer to understanding one universal component of the condition regulating the availability of ellipsis. In closing, I discuss the long-term prospects of a collaborative research program focusing on silent expressions in Mayan, a novel empirical domain in an otherwise well-described language family (Aissen et al. 2017).